In crowded visual scenes, attention is needed to selectively process information from relevant stimuli and to filter out irrelevant distracters. Accordingly, studies in humans and animals have shown that neurons in several different brain structures show enhanced responses when attention is directed into their receptive fields. However, there is still little understanding of how these ubiquitous attentional effects are actually generated through the interactions among these structures, particularly when the attentional cues arise from cognitive factors and task demands. To design an effective neural prosthesis or to treat people with attentional disorders, we need a better understanding of attention at the systems level. Three key structures thought to provide feedback to visual cortex that mediates attentional effects are the prefrontal cortex, the posterior parietal cortex and the pulvinar. In the planned experiments, we will compare the roles of each of these three structures in providing feedback to area V4 in visual cortex. Area V4 plays a central role in the relay of visual information along the ventral stream that underlies object recognition, and we have previously shown that V4 neuronal responses are modulated by attention and that damage to V4 causes attentional impairments. We have also recently shown that neurons in prefrontal cortex have attentional latencies that are short enough to mediate some of the attentional effects on V4 neuronal responses, and that prefrontal neurons have coherent activity with cells in V4. This coherent activity is time-shifted across a wide range of frequencies by about 10 ms, which may be the critical time to allow for functional interactions. In Aim 1, we will add recordings from the posterior parietal cortex to the prefrontal recordings, to compare the roles of these two structures in modulating activity in V4. The pulvinar provides an alternative anatomical route for signals from prefrontal and parietal cortex to influence V4. Therefore, in Aim 2, we will record simultaneously in the pulvinar and area V4, to test whether pulvinar neuronal properties are consistent with this feedback role. However, neurophysiological recordings alone can only provide evidence for correlations in activity, not for causality. Therefore, in Aim 3, we will supplement the neural recordings with suppression of activity in prefrontal and parietal cortex and pulvinar, to test causal hypotheses about the role of feedback from each structure in modulating V4 responses. For these experiments, we will use techniques that we and our collaborators have recently developed for the optogenetic suppression of neural activity using the proton pump, Arch-T, which is delivered by lentivirus. With Arch-T we will be able to suppress activity with resolution in the tens of milliseconds, at critical time points, to gain new mechanistic insights into the feedback to V4. In total, we expect these studies to give us the best account so far of how the interactions among multiple brain structures leads to effective visual processing with attention.